MESOSAUKIA OF SOUTH AFKICA. 699 



known to indicate that it is a fragment of the coracoid of an allied 

 genus, in which there was no squamous overlap of the bones. The 

 British Museum examples of Stereosternum show that the cervical 

 vertebrae are very short. They are exposed laterally. The centrum 

 is badly preserved. The neural spine is long, thin, flat ; with the 

 anterior and posterior borders sub-parallel, slightly converging 

 superiorly. The extremity of the spine is truncated. The neural 

 spines are inclined backward, and increase in length towards the 

 dorsal region. The neurapophyses are constricted from back to 

 front, so as to cause the anterior angle to extend forward as a pre- 

 zygapophysis, and there is a strong post-zygapophysis with a semi- 

 circular excavation beneath it for the intervertebral nerve. There 

 are six vertebrae preserved anterior to the humerus. In the Cape 

 Town Mesosaurus these would extend forward to the middle of the 

 cervical region.^ 



The measurement from the humerus to the femur is 13*5 centi- 

 metres, and in this length are twenty vertebrae. The transverse 

 measurement over the middle of the dorsal ribs is 4-2 centimetres. 

 There are twenty pre-sacral vertebrae which bear dorsal ribs. The 

 early vertebrae of the dorsal series have the centrum short, measuring 

 I centimetre. The under side of the centrum is wide anteriorly, 

 convex from side to side, and less convex from back to front. In 

 the specimen R. 537, twenty dorsal vertebrae measure 16*2 centi- 

 metres ; that example is therefore somewhat larger. There are no 

 lumbar vertebrae in Stereosternum. 



The early dorsal ribs are slender proximally, and, enlarged, they 

 are club-shaped distally. In R. 537 



the ribs appear to be flatter, but Fig. 2.— Sacrum and ilium 

 this may be the eff"ect of com- of Stereosternum. 



pression. 



In the specimen E. 536 there 

 are four sacral vertebrae charac- 

 terized by their sacral ribs, sup- 

 porting the ilium. Probably only 

 the last two are rightly accounted 

 sacral, and the anterior, which have 

 the ribs converging outward, may 

 be sacro-lumbar. There is no 

 reason for believing that any of the 

 sacral vertebrae are anchylosed. 

 In Giirich's Ditrocliosaurus from Eestored. 



Hope Town the sacral vertebrae ., ... ^ . , 



, -i ii. J mu • 14-1, w> ilium ; a, acetabulum, 



appear to be scattered. Ine width 



of the stronger transverse processes which are opposite the aceta- 

 bulum in Stereosternum is about | centimetre. 



The early caudal vertebrae of Stereosternum seen from above show 

 the lateral notching behind the transverse processes, which defines the 

 post-zygapophyses. The transverse processes of the caudal vertebrae 



^ Prof. Cope records nine cervical vertebrae besides the atlas, ' Am. Nat.' vol. 

 xxi. p. 1109. 



